*NOTICE: This is the peer reviewed version of the following paper:

Bossetta, M. (2017). Fighting Fire with Fire: Mainstream Adoption of the Populist Political Style in the 2014 Europe Debates between Nick Clegg and Nigel Farage. British Journal of Politics and International Relations, 19(4), 715-734.

Published at: https://doi.org/10.1177/1369148117715646

Fighting Fire with Fire: Mainstream Adoption of the Populist Political Style in the 2014 Europe Debates between Nick Clegg and Nigel Farage

Michael Bossetta (University of Copenhagen)

AbstractAdvancing the concept of populism as a political style, this study compares the debate

performances of two British party leaders, Nick Clegg and Nigel Farage, as they clashed in a

pair of televised debates over Britain’s EU membership ahead of the 2014 European

Parliament elections. The argument tested is that under certain conditions, mainstream

politicians will adopt a populist communication style while retaining a non-populist agenda.

A mixed methods approach combines computational text analysis with a qualitative rhetorical

analysis to demonstrate how the populist and non-populist style can be distinguished and

compared systematically. The results suggest that Clegg, although maintaining a non-populist

ideology, adopts features of the populist style after losing the first debate. Farage’s

communication style, conversely, remains stable to the point of statistical significance. This

suggests that one explanatory factor of populists’ success is the consistency of their message

and rhetorical delivery, bolstering their perceived authenticity among voters.

Introduction Across Europe, the increase in electoral support for so-called populist parties is dazzling. So,

too, are the politicians who comprise them. Conflictual, controversial, and crude, the new

wave of populist challengers is a far cry from the deliberative, temperate, and polished

politicians that have typically governed advanced liberal democracies. Although the

academic literature tends to stress the similarity of populists by focusing on their ideological

affinities (Mudde 2004; Stanley 2008; Rooduijn et al. 2012; Kriesi and Pappas 2015),

European populist actors often differ along their domestic policy agendas, position on the

political spectrum, and degree of opposition to European integration. An illustrative example

of this difference is the present state of right-wing populist parties in the European Parliament

(EP), where they resemble more of a motley crew than a traditional party family. Despite

garnering unprecedented electoral support in the 2014 EP elections by campaigning against

both the national and European political establishment, right-wing populists cannot seem to

coalesce into a cohesive front against the political mainstream at the European level. The UK

Independence Party, the Danish People’s Party, and the Front National – each winning the

2014 EP elections within their respective domestic arenas – are currently split among three

distinct political groups in the EP. What the populist firebrands seem to share more than

ideological cohesiveness is a certain style of political communication (Jagers and Walgrave

2007; Moffitt 2016).

The present study applies the concept of populism as a political style by comparing the

debate performances of a paradigmatic populist, Nigel Farage, directly against those of an

established politician, Nick Clegg, as they clashed in a pair of televised debates over Britain’s

EU membership before the 2014 EP elections. The Europe debates constitute only the second

set of party leader debates in British history and the first time since 1975 a political debate

over EU membership was broadcast live to the national public. The head-to-head

confrontation between the two party leaders offers an interesting case to examine how

diverging policy positions on EU membership were argued before the electorate two years

prior to the historic 2016 Brexit referendum. Additionally, the juxtaposition of a high-ranking

member of government (Clegg) and a populist challenger (Farage) mimics wider power

2

struggles across the continent, where mainstream politicians are increasingly pressured to

engage unconventional, populist politicians over key policy issues regarding immigration and

national sovereignty. The present study takes the Europe debates as an empirical site to

investigate the discursive content and rhetorical strategies of the party leaders in order to

answer the research question: Do mainstream politicians adopt a populist political style while

maintaining a non-populist agenda?

Scholarly attention devoted to how mainstream political actors engage populists is relatively

scant and predominantly focuses on whether mainstream parties adopt the political agenda of

populists on issues like multiculturalism, immigration, or European integration (Bale et al.

2010, van Spanje 2010; Akkerman 2012; De Lange 2012; Mudde 2013; Schumacher and van

Kersbergen 2014). Taken together, these studies suggest that populists’ electoral success can

motivate both left- and right-wing mainstream parties to enact policy shifts, especially

regarding social and cultural policies. They also stress, however, that the mainstream’s co-

option of populists’ policies varies across cases and is largely contingent upon the specific

conditions of the political system, such as whether the mainstream party is in government or

opposition, the bargaining power of populists in forming coalitions, and the mainstream

parties’ willingness to collaborate with the populist insurgency.

Motivated to test whether populists can also impact the practices of the political mainstream

at the level of individual politicians, this study takes an empirical case where a populist and

non-populist party leader publically debated dichotomous policy positions on EU

membership. Farage, the populist case, strongly advocated a British withdrawal from the

Union whereas Clegg, the non-populist case and Deputy Prime Minister, unwaveringly

supported Britain to remain in. The study’s overarching claim is that under certain conditions,

established politicians may adopt a populist style of political communication while retaining a

non-populist agenda. The populist style is construed broadly as patterned practices of

3

political communication seeking to promote conflict and convey a sense of crisis. Although

politicians in power would rationally attempt to avert political conflict and stem perceptions

of crisis, the increasing electoral success of populist challengers places pressure on

mainstream politicians to ‘fight fire with fire’ and adopt a populist style. By changing their

political communication strategies but not necessarily their political agendas, mainstream

politicians can advocate their existing policy platforms while attempting to appeal to a

broader voter base.

The incentive for mainstream political actors to alter their communication strategy arises out

of several challenges currently facing many traditional parties: declining party membership,

low trust in incumbent governments, the rising popularity of anti-establishment parties, and

the propensity of the media to cover populist politicians (Mazzoleni 2014). Considering these

challenges, a number of conditions can incentivize mainstream adoption of the populist style,

such as decreasing poll numbers, increasing public support for populist parties, and the

saliency of polarizing issues in media coverage. These conditions are not investigated here

but are taken as the constant environment in which to test the difference in political styles

between Farage and Clegg. To answer whether mainstream politicians adopt a populist style

under these conditions, two hypotheses are developed. The first examines whether the

populist style can be empirically measured via the party leaders’ discourse; the second seeks

to expound any change in their political style across the two debates.

After theoretically developing the concept of political style, a mixed methods research design

is introduced to operationalize the political styles of Farage and Clegg. Quantitative,

computational text mining methods are complemented with a qualitative, rhetorical analysis

to assess both the discursive content and rhetorical strategies used by the politicians in

arguing their positions on EU membership. The results suggest that the styles of the two

politicians are indeed distinct, and that Clegg, having lost the first debate, adopts traits

4

characteristic of the populist style in an attempt to win the second. By contrast, Farage’s

discourse and rhetorical appeals remain remarkably stable across both debates. This suggests

that one of the underlying mechanisms explaining populists’ success is the consistency of an

anti-establishment message. The consistency of Farage’s political communication is

interpreted as bolstering his authenticity among parts of the electorate, who sanction him as a

legitimate challenger to the political establishment.

Populism as political style The academic community continues grapple with arriving at a sufficiently agreeable

consensus over populism’s definition. Some scholars advocate that populism is best

conceptualized as a ‘thin-ideology’ (Mudde 2004) that reduces the body politic to two groups

(i.e., people and elite), posits an antagonistic relationship between them, and affirms that the

‘people’ hold the sovereign right to select their governors. Others argue that this

understanding of populism is too catch-all and prefer to focus on the specific discursive

(Aslanadis 2015) or performative (Moffitt and Tormey 2013) elements of populists’

communication. According to Kriesi (2014, 364), neither of these approaches is necessarily at

odds with populism’s ideological definition, since ‘the populist ideology manifests itself in

the political communication of populist leaders.’

However, arguing that populist communication emanates from a populist ideology poses the

implicit risk of selection bias. If anti-elitism is a defining feature of the populist ideology and

is operationalized as critical discourse against the political mainstream, then empirical

measures of populism will, almost by definition, register political outsiders as ‘more populist’

than established elites. People-centrism, another commonly ascribed component of the

populist ideology, is not a sufficient counter-weight to correct this bias. People-centrism is

the fundamental tenant of all representative politics, and appealing to ‘the people’ – a socially

5

constructed entity (Laclau 2005, 72-74) – is not a particular feature of populists. One could

even envision a situation where an established political actor denounces a populist challenger

by stating: ‘We, the tolerant people of Britain, reject your illiberal principles and deem you

illegitimate to govern.’ If issued by a populist to a mainstream party leader, this statement

would classify as populist on account of its appeal to the people and implicit criticism of the

elite. If uttered, however, by an established politician to a populist in power, would the same

phrase still be populist? And if so, would it indicate a populist ideology?

Aiming to recast populism as a concept applicable to more than just the fringes of a political

system, the present study advocates approaching populism as a political style. Following

Moffitt (2016, 28-29), a political style is understood as ‘the repertoires of embodied,

symbolically mediated performance made to audiences that are used to create and navigate

the fields of power that comprise the political.’ Put another way, a political style refers to the

empirically observable, patterned practices of communication exhibited by politicians in

promoting their political agenda, both in terms of discursive content and performative

features accompanying its delivery. Such performative features include the rhetorical

strategies used in making an argument, non-verbal cues like body language and dress, and

properties of language like accents, prosody, and diction. While performative features may be

‘secondary’ (Aslanidis 2015, 11) to discursive content in identifying populism, they are

intertwined with the communication strategies of political leaders and are thus integral to

explaining why populist messages are successful. The tendency of scholars to measure the

degree of populist discourse in political speeches or party manifestos seems to imply that

populists’ success is reducible to words or framing strategies – which is unlikely. Rather,

systematic empirical inquiry into exactly how populist actors attempt to make their messages

persuasive can help identify the elusive ‘micromechanisms’ (Pappas 2016, 14) that explain

6

populists’ resonance with a given constituency and further, contribute to constructive theory

building.

What unites populists, fundamentally, is their criticism of the political status-quo.

Approaching populism as a political style means investigating how political actors convey

this criticism convincingly to the electorate, influencing their political preferences and

mobilizing participation around alternative political platforms. The present study contends

that rhetoric, the art of persuasion, is a critical component of political communication that,

although underutilized in the populist literature, can inform our knowledge of how populist

interventions are formulated, achieved, and maintained. The connection between populism

and rhetoric is not new (Canovan 1984; Kazin 1998) but has been rebuffed for lacking

analytical clarity, since rhetoric is part-and-parcel of all political communication. However,

the argument put forth here is that the systematic deconstruction of politician’s rhetorical

strategies can reveal patterns of communication that help explain populists’ success.

Approaching political communication from a rhetorical lens acknowledges the strategic,

agential, and creative activity underpinning the process of formulating – and performing –

political arguments (Finlayson and Martin 2008; Martin 2015).

The study of rhetoric in British politics has illuminated five aspects of political

communication relevant for understanding populism as a political style. First, rhetorical

strategies can be utilized to bring about change in the political status-quo, and one pattern that

emerges from the literature is political actors’ use of rhetoric to establish legitimacy by

advocating a ‘new’ form of politics. Atkins (2011), for example, demonstrates how the

Labour Party utilized moral arguments to rebrand itself as ‘New Labour’ and wrest power

from the Conservatives. Similarly, Nick Clegg and the Liberal Democrats were able to

challenge the establishment in 2010 by promoting a platform of ‘new politics’ (Parry and

Richardson 2011, 480), a progressive mantra that was later rhetorically crafted to bind the

7

Liberal Democrat/Conservative coalition together, despite their ideological differences

(Crines 2013a, 212).

Second, the literature on British rhetoric stresses the importance of politicians’ personality

and style in communicating their political agendas. Recent anthologies of Labour (Crines and

Hayton 2016) and Conservative (Hayton and Crines 2015, 198) politicians highlight how

much of their success ‘depends on the personality and style of each individual orator’. Third

and related, both old and new media increasingly influence the communication strategies of

politicians. Complementing the case made by Atkins, Charteris-Black (2011, 225) argues that

much of New Labour’s success is attributable to the communication style of Tony Blair, who

was able to harness the power of metaphor to convey messages in a way fitting the media’s

penchant for ‘brevity, clarity, and simplicity.’ Both David Cameron and Nick Clegg shared

Blair’s talent for addressing the media, and Charteris-Black (2011, 20) accredits their

‘relaxed and informal style’ as contributing to their persuasiveness in the 2010 party leader

debates. Rolfe’s (2016) work demonstrates how political actors can use digital media to craft

the argument for a ‘new politics’, and he argues that both Barack Obama and Julian Assange

shared a populist rhetoric complimented by their innovative use of digital technologies.

Similar practices are observed by populist party supporters on social media; a recent study of

citizens’ Twitter use during the 2015 British General Election finds UKIP supporters

overwhelmingly active in issuing mobilizing calls for action (Dutceac Segesten and Bossetta,

2016).

Lastly, the study of British political rhetoric has revealed a general shift in political

communication towards a simplified, populist discourse. While Dommett (2014, 83) argues

that Labour’s brass have adopted the rhetorical technique of “using everyday language” from

the Blue Labour movement, Finlayson (2012, 759) notes “an increasing tendency across all

parties to cite ‘ordinary’ people” through the strategic use of anecdote (Atkins and Finlayson,

8

2013). The prevalence of invoking ‘the people’, witnessed even at the regional level of Welsh

politics (Moon, 2014), indicates a stylistic change in British political communication culture

emboldening the claim that the “very idea of Britishness is negotiated through rhetoric”

(Atkins and Turnbull 2014, 173). The present study contributes to this body of literature by

providing an analytical framework to identify the populist style through rhetoric.

Complimenting this theoretical contribution, the study provides a methodology grounded in

the interpretive approach used by British rhetoricians while adding a systematic, quantitative

component that captures the populist style at the level of language and performance.

In order to delineate the particular features of the populist style, the study builds on Moffitt’s

(2016, 44) theoretical model distinguishing the populist style from the technocratic (or non-

populist) style. Both styles are distinct modes of garnering political legitimacy and reflect

dichotomous, ideal types of representation. At one end of the spectrum, adherers of the

populist style appeal to the people versus an elite, exhibit bad manners, and strive to perform

a sense of crisis. At the other end, politicians enacting a technocratic style will appeal to

expertise or scientific knowledge, exhibit good manners, and aim to promote a sense of

stability or measured progress (Moffitt 2016, 26).

Figure 1: The Technocratic-Populist Political Style Spectrum (Moffitt, 2016: 44)

In a first step towards operationalizing political style, two hypotheses are formulated to test

the enactment of the populist and technocratic style by Nick Clegg and Nigel Farage, who by-

and-large correspond to the model’s ideal types. Discussed in further detail below, Clegg is a

well-mannered, polished politician who led the Liberal Democrats into a coalition

9

government with the Conservatives in 2010 by running a platform of ‘honest’ and ‘sincere’

politics. Farage, by contrast, is the leader of the populist UK Independence party (UKIP) and

is often depicted as a politically incorrect politician, renowned for his bombastic speeches in

the European Parliament. Taking the 2014 Europe debates between the two leaders as a case

study, the first hypothesis states that:

H1: Farage will exhibit a populist style and Clegg will exhibit a technocratic style.

More specifically, three individual features of the populist style are examined against those of

the technocratic style in an exploratory attempt to apply Moffitt’s model. Moffitt marks a

relevant stylistic distinction between good and bad manners. The populist style incorporates

simplistic language, colloquial parlance, and unconventional behavior relative to the norms of

a given political culture. The technocratic style, on the other hand, employs scientific

language, esoteric parlance, and respects political customs and conventions. Although speech

and behavior are distinguishable, the two can be also be considered complementary, since

formal speech is unlikely to be accompanied by inappropriate behavior and vice versa. As

speech is more easily operationalized than behavior, I focus on the former and hypothesize in

H1a that:

H1a: Clegg will use more formal language than Farage.

Another aspect in Moffitt’s model refers to the narrative – or ‘script’ (Alexander 2011, 57-

59) – that political actors choose to communicate. Populists aim to convincingly portray

society as in a state of crisis (Rooduijn 2014; Moffitt 2015), whereas technocrats usually

attempt to counter the perception of crisis by conveying stability or progress. As Hay (1996)

demonstrates, crisis is a discursive construction. The mediation of crisis versus stability, then,

refers largely to the content of political discourse and the ‘story’ a political actor tries to tell.

Hypothesis H1b posits that:

10

H1b: Farage will seek to discursively mediate a sense of crisis, whereas Clegg will mediate stability regarding Britain’s EU membership.

One of the primary strengths of Moffitt’s model is that it uncouples populism from the

ideological connotations that have long been associated with the concept. However, the first

dimension of Moffitt’s model, ‘appeals to the people versus the elite,’ arguably excludes

those in power from being considered to enact a populist style, since governing elites will

never rationally attack themselves. By the same token, populist politicians may refer to their

expertise outside of politics in order to establish credibility among voters. To avoid this bias,

hypothesis H1c proposes a focus on the rhetorical appeals politicians utilize in making their

arguments persuasive: ethos, logos, and pathos. These appeals, grounded in the work of

Aristotle, have been the backbone of other analyses of British political rhetoric (Atkins et al.

2014, 6-8; Martin 2014, 57-65).

The rhetorical appeal to ethos primarily relates to legitimizing a speaker’s authority

(Finlayson 2012, 760) and is directly related to democratic representation and power. Ethos

strategies establish a speaker’s credibility among an audience, usually through references to

one’s own personal character, qualities, or accomplishments. If a successful political

performance depends on the extent to which an actor achieves ‘authenticity’ by coming

across as ‘straightforward, truthful, and sincere’ (Alexander 2011, 7), then ethos is a

necessary condition for a political leader in making successful representative claims.

Logos appeals can be identified when an argument is made on the basis of facts, examples, or

reasoning. While voters need access to factual information in order to form accurate political

preferences and make informed choices at the polls, political actors attempt to cue public

opinion by selectively presenting information to voters (Hooghe and Marks 2005). Moreover,

what constitutes a ‘fact’ can be based on false or tenuous premises. This notwithstanding,

logos is evidenced by the invocation of facts, irrespective of their verifiability.

11

The third rhetorical proof, pathos, refers to making an argument by appealing to the

sentiments of the audience. Emotional arguments can be either positive or negative, with the

former generally promoting trust and the latter inducing a sense of threat. Affective messages

are powerful tools in the repertoires of political actors, since emotional arguments can be

seen as unfalsifiable and have been demonstrated to be more effective than cognitive-based

appeals (Marcus 2000, 232).

The three rhetorical appeals are not mutually exclusive. They are often combined for

persuasive effect and tailored to the context of a particular speech occasion. The present

study adds to the established three appeals a fourth, denigration, to refer to the direct or

indirect slander of a political opponent. Denigration is antithetical to ethos. Instead of seeking

to build credibility through ethos, a political actor can attempt to undermine an opponent’s

legitimacy via attacks on his or her credentials, arguments, or personal character. Ultimately,

the goal of denigration is to achieve an effect similar to what Krebs and Jackson (2007, 36)

refer to as ‘rhetorical coercion,’ when an opponent lacks a ‘socially sustainable rebuttal’ and

is ‘compelled to endorse a stance they would otherwise reject.’

Relating these four rhetorical appeals to the concept of political style, this study argues that

the populist style aims to establish political legitimacy via appeals to pathos while also

striving to successfully denigrate a constructed enemy. The connection between populism and

pathos has been acknowledged previously (Reisigl 2008, 103), and the denigration of a

political elite or immigrant other, usually through blame attribution (Vasilopolou et al. 2014),

is a necessary foil to construct the identity of ‘the people.’ The technocratic style, on the other

hand, seeks to maintain legitimacy and reach consensus via appealing to personal expertise

(ethos) and supported facts (logos). Crines (2013b, 84) has previously described the rhetoric

of the British mainstream as ‘logos-driven’ and ‘deliberative.’ Whereas the technocratic style

seeks to convey information substantiated by appeals to personal or expert authority, the

12

populist style seeks to drive conflict via appealing to emotions. Based on this nuanced

understanding of the populist style, the last expectation for H1 is that:

H1c: Farage will argue his anti-EU case using appeals to pathos and denigration, while Clegg will argue the pro-EU case making appeals to ethos and logos.

Whereas Hypothesis 1 tests whether the discourse and rhetoric of the two leaders corresponds

to their expected political styles, Hypothesis 2 is structured to identify changes in political

style across the two debates. While both debates fulfill the conditions propitious for

mainstream adoption of the populist style (low poll numbers for established politicians, rising

popularity of populists, and a polarizing issue salient in the media), in the second debate these

conditions are more pronounced. Clegg was interpreted as losing the first debate (YouGov

2014a), Farage was seen as having won, and although the polarizing issue was the same

across the debates, the second debate was hosted by a larger media outlet (BBC) than the first

(LBC). Expecting the circumstances in the second debate to be more favorable towards

adopting a populist style, Hypothesis 2 expects that:

H2: Clegg’s performance will be more populist in the second debate compared to the first.

The next section outlines the context of the empirical case chosen for the study, with an

emphasis on why the case fits the conditions posited to incentivize mainstream adoption of

the populist style.

Case SelectionPolitical debates are classic bouts of rhetoric where politicians aim to convey their political

messages convincingly to the electorate. Debates are therefore primary loci to capture

political styles in action, and the cases selected to test the populist and technocratic styles are

the 2014 Europe debates between Nigel Farage and Nick Clegg. The debates comprise only

the second set of party leader debates in British history and the first time since 1975 a public

13

political debate over Britain’s EU membership was broadcast live to the national public. The

debates’ context fulfills the criteria posited to incentivize the adoption of a populist style by a

mainstream politician: the mainstream politician is unpopular and polling poorly, the populist

politician is polling well, and the topic of the debates (EU membership) is highly polarized

and salient in the media.

The once marginal UKIP, spearheaded by Farage, gained public support by promoting a

populist, anti-establishment narrative framing national political parties, multinational

corporations, and EU politicians as corrupt elites stripping the national sovereignty away

from the British people. Although the anti-EU position is not monopolized by UKIP, their

version of Euroscepticism is argued to be populist in comparison to the ‘technocratic, more

elite-based Euroscepticism’ (Tournier-Sol 2015, 144, my emphasis) exhibited by the

mainstream Conservative Party. At the time of the debates, most polls projected UKIP would

garner around 30% of the popular vote for the EP elections, placing the party in strong

contention to emerge as the election’s victor.

On the other side of the EU cleavage, the LibDems were the sole staunchly pro-EU party in

Britain. Although generally a small party on the fringes of a British political system, in 2014

the LibDems were the minority party in Britain’s first coalition government since the Second

World War. The historic arrival of a third party alternative to a traditionally two-party system

was due in part to Clegg’s highly successful debate performances in Britain’s first televised

leader debates before the 2010 national elections (Parry and Richardson 2011). However,

Clegg’s star power eroded quickly after reversing a number of campaign promises,

particularly not to raise tuition fees, and at the time of the debates the LibDems were polling

an abysmal 10% for the EP elections.

14

In an attempt to bolster the LibDem’s poll numbers ahead of the elections, Clegg challenged

Farage to an open, public debate over Britain’s EU membership. The first debate, taking

place on March 26th, 2014, was transmitted over the radio, streamed live on YouTube, and

televised on Sky News. The second debate, held two weeks later on April 2nd, 2014, was

broadcast live exclusively on the public broadcast station BBC Two.

The Europe debates are a particularly suitable case since the two politicians involved fit by-

and-large with the populist/technocrat spectrum. An outspoken anti-populist, Clegg promotes

a pro-EU position, ascribes to a left-wing liberal ideology, and speaks from a high profile

position in government. Farage, generally regarded as a populist, advocates an anti-EU

position undergirded by a right-wing, anti-establishment ideology and speaks from a position

of government opposition. Moreover, the format of the debate provides laboratory-like

conditions to longitudinally test their communication strategies. Despite being hosted by

different media outlets, the debates were both on the same topic (EU membership) and

maintained identical formats. Each candidate had one minute for opening and closing

statements, and in between them the candidates fielded pre-screened questions directly from

the audience. A journalist moderated the discussions and interjected clarifying questions.

Both debates were filmed live before a studio audience that was selected by the independent

polling organization ICM to represent an equal number of ‘In’ and ‘Out’ supporters, as well

as a smaller number of undecided voters.

MethodTo compare the political styles of Clegg and Farage across the two debates, a sequential

mixed methods approach is employed and divided into two phases. The first consists of a

series of quantitative, computational text mining methods based on automated content

analysis, natural language processing, and statistical clustering. The strength of the

computational approach is that it compares the two styles in strictly similar ways while

15

presenting the general picture of the two politicians’ messages. While previous work on

populism (Rooduijn and Pauwels 2011, 1279) highlights ‘accuracy’ and ‘reliability’ as key

advantages of computer-based content analysis, they also advocate ‘an integrated man-

machine approach that filters out the disadvantages and emphasizes the strong points’ of both

methods. Therefore, the second phase deepens the understanding of the differences between

the two styles by using a qualitative coding of rhetorical appeals. A heuristic, qualitative

reading has the advantage of being sensitive to the particular ways ideas are structured and

expressed through speech by taking into account context largely neglected in quantitative

methods. The same sample is used in both the computational and the qualitative phases to

increase the validity of the results (Creswell 2014, 225).

The data collection process consists of transcribing the debates from available video

recordings archived on YouTube ad litteram. This means that incomplete sentences,

crosstalk, and emotional instances of interruption, such as laughter (‘ha’) or other

interjections (‘aw’ or ‘no’), are included in the transcription.

The quantitative phase is divided into five steps using the ‘quantitative discourse analysis

package,’ or ‘qdap’ (Rinker 2013), for the programming software R. In the first step, which is

aimed to test Hypothesis 1a and assess the formality of language used by each politician, I

use the formality function of qdap. Based on a formula developed by Heylighen and Dewaele

(2002), the frequency of parts of speech that generally carry substantive information are

counted and divided by those that typically convey information implicitly through context.

This natural language processing method corresponds to the linguistic differences posited by

the populist/technocrat spectrum:

‘A formal [technocratic] style of expression is characterized by detachment, precision, and objectivity, but also rigidity and cognitive load; a contextual [populist] style is much lighter in form, more flexible and involved, but correspondingly more subjective, less accurate and less informative’ (Heylighen and Dewaele 2002, 334).

16

To test Hypothesis H1b, that Farage will mediate a sense of crisis and Clegg will attempt to

convey stability by staying in the EU, I perform a word frequency analysis that singles out

the top 15 words used by each politician per debate. This method identifies the most frequent

topics discussed. In a third step, I identify the top 10 words used exclusively by each

politician, in order to expound the idiosyncratic strategies in laying out their respective

arguments.

I operationalize Hypothesis 1c, which tests the rhetorical appeals of the politicians’ discourse,

through a qualitative reading of the text aided by the coding software MAXQDA. Delineating

rhetorical appeals is a complicated exercise since different appeals are often combined and

interwoven to enhance persuasive effect. Thus, to perform the qualitative reading as

systematically as possible, a deductive coding scheme is developed and applied to a thematic

unit of analysis (Budd et al. 1967, 34). I choose not to use the sentence as the unit of analysis,

since sentences were difficult to separate in the spoken discourse of the politicians, especially

during instances of crosstalk.

While there are no existing guidelines regarding how to systematically operationalize

rhetorical appeals, the coding scheme is structured as follows. Ethos was coded to when the

speaker specifically mentioned personal accomplishments or previous professional

experience; both are considered means to garner legitimacy and establish credibility. Pathos

was coded when values, emotional language, or interests relating to the people (e.g. job

security, immigration, or safety against criminals) were invoked. Pathos also includes general

identitarian references to ‘we’ or ‘us’ at the national level, such as ‘our place in the world’ or

‘let’s not meddle in international affairs,’ since these can be seen as implicit references to the

audience. Logos was coded when the speaker made a reference to a fact, figure, or concrete

empirical example. This could be performed either through directly quoting an authority,

report, or through mentioning an unsupported but reasonable figure, like ‘we export a million

17

cars per year.’ Lastly, denigration was coded when one opponent slandered the other either

directly, e.g. ‘you are willfully lying to the British people,’ or through indirect slights like

‘what you are owed are the facts, not simply a lot of opinion,’ which are considered not

appeals to logos as much as attempts to attack an opponent. If no specific appeal could be

identified, for example when an argument was not explicitly being made or during instances

of uninterpretable crosstalk, the text was coded as Other. Following the first coding, one-

quarter of the data was recoded after a three-month period and resulted in a coder reliability

score of 77%.

Hypothesis 2, pertaining to the change in styles across the debates, is tested qualitatively

through the rhetorical analysis but also quantitatively via statistical clustering methods that

compare the difference in word frequencies across each of the variables (in this case,

politician and debate). The fourth step of the quantitative phase uses multidimensional

scaling (MDS), which graphically represents the politicians’ discursive content against one

another and across the debates. Lastly, I use agglomerative hierarchical clustering (Kaufman

and Rousseeuw 2005, 44) as a compliment to the MDS model. Whereas MDS is primarily

used to uncover differences among variables, hierarchical clustering attempts to identify

similarities in the data by clustering the variables into groups. The R package ‘pvclust’

(Suzuki and Shimodaira 2015) is also used to quantify the strength of correlations between

the clustered texts.

18

Results

Figure 2: Formality Scores

19

Figure 2 presents the formality scores for each politician, as well as the moderators for

comparison. The higher the formality score, the more precise and informative one’s speech

is; the lower the formality score, the more contextual and less informative one’s speech is.

Figure 2 shows that the two moderators have the highest formality scores, which is to be

expected given their role of relaying substantive questions from the audience and asking

clarifying questions to the politicians. Farage’s first (LBC) debate is shown to be the most

formal of the two party leaders, while his second (BBC) debate is the least formal. Farage’s

particularly formal language in the LBC debate may be explained by the fact that this was his

first debate appearance alongside another national party leader, and that he was attempting to

attain legitimacy by presenting his platform pointedly. Another reason may relate to Clegg’s

demonstrated use of ‘demotic modes of expression and direct address to audiences’ in

previous televised debates (Washbourne 2013, 113). Clegg’s formality scores are almost

identical across the two debates but not significantly higher than Farage’s. Hypothesis 1a,

that Clegg’s language would be more formal than Farage’s, is largely unconfirmed.

Nick Clegg Nigel Farage LBC (n=5262) BBC (n=5112) LBC (n=4488) BBC (n=4971) Word Freq Word Freq Word Freq Word Freq

people 42 people 48 people 33 Nick 33 European Union 39 Nigel Farage 35 Nick 24 people 30 country 28 European Union 29 European 20 trade 25 job(s) 37 world 26 Europe 19 European Union 19 British 23 country 22 years 18 country 18 million 22 Britain 19 British 17 let's 18 rules 22 British 17 country 17 world 18 actually 21 percent 16 European Union 17 dear 16 law(s) 18 change 15 million 17 European 16 fact(s) 17 million 15 actually 16 percent 16 Europe 17 simply 15 law(s) 15 British 14 simply 17 actually 14 hundred 14 Europe 14 thousand 17 countries 14 trade 14 own 13 European 16 European 14 question 13 years 12 Nigel Farage 16 modern 14 Britain 12 immigration 11

Table 1: Word Frequency Analysis Result

20

Table 1 presents the results of the word frequency analysis, which is aimed at assessing the

discursive construction of crisis versus stability. The columns divide the most commonly

used words each by politician and debate. For both Clegg and Farage, people is the most used

word in nearly all the texts. This is considered not an indicator of populism since people is

among the top two hundred most commonly used words in English (Fry, 1997). Words used

prominently by both politicians in each case are European Union, Europe, European,

country, British, and million. The presence of these words is rather unsurprising given the

debates’ theme and format, inviting topical discussions about the EU and the UK supported

by figures, as indicated by million.

Looking only at Clegg, references to job(s) occur 37 times in the LBC debate but does not

make the top 15 words in the BBC debate (used only 12 times there). Other words present in

the top 15 list for Clegg in the LBC debate are: law(s), fact(s), rules, and thousand. These

words suggest argumentation based on facts and figures, and they are largely absent in the

BBC debate. Instead, mentions of Nigel Farage more than double, suggesting an increased

focus on Farage’s person from the first to second debate. This personalization comes at the

expense of substantive words like facts and laws, while at the same time words indicating

figures, like million and thousand, decrease. Clegg’s BBC list also hints to a heightened

domestic focus, with the words Britain and British appearing 36 times combined, and the

total number of references to the EU and Europe decreasing in relation to the LBC list.

For Farage, mentions of the EU are roughly equivalent to those of country, and the

appearance of trade and years on both lists suggest that trade and historical references are

key components to Farage’s arguments. Like in Clegg’s lists, words that indicate facts and

figures, such as million and hundred, are present in the first but absent from the second

debate. Immigration, a key UKIP policy issue, is mentioned 11 times in the BBC debate

(compared to 6 times in the LBC debate).

21

All in all, the word frequencies suggest that the degree of deliberation, approximated through

mentions of facts and figures, decreased from the first to the second debate, giving way to

more mentions of the opponent for Clegg and increased focus on trade and immigration by

Farage. Table 2 presents the unique words used by each politician in the debates, in order to

gain a better idea of idiosyncratic arguments of the two politicians.

Table 2: Word Difference Lists

While some words that appear in the word frequency lists are also present in Table 1, a

number of others that are relevant to the analysis come forth. Looking at Clegg’s LBC list,

we see the word elsewhere (a framing strategy to euphemize the word ‘immigrants’ as

‘people from elsewhere’), facts, clout (referring to Britain’s international clout in being an

EU member), and pull (in relation to Clegg’s chosen metaphor of leaving the EU, namely to

‘pull up the drawbridge’). For Clegg’s BBC list, the word dangerous is particularly high

(referring to the ‘dangerous fantasy’ that leaving the EU will help Britain), alongside clock

(another metaphor that compares leaving the EU to ‘turning the clock back’), climate

(referring to climate change), and Vladimir Putin (in reference to an interview held between

the debates where Farage claimed to admire the Russian president).

Words unique to Farage in the LBC debate include framing the European Union as Brussels,

bringing up the cost of EU membership, and deploring the free movement of people that he

describes as an open door, which has left the British people lacking the ability to control their

22

border. The word best refers to UKIP’s slogan (‘The best people to govern Britain are the

British people themselves’), while mister refers to his mentions to other party leaders, namely

David Cameron of the Conservatives and Ed Miliband of Labour. In the BBC debate, many

of the same words are mentioned but with increased frequency (control, free, and door), and

the EU is often referenced as a political union. Some words, like dear (as in, ‘oh dear’) and

ha (indicating laughter), point to particular patterns in Farage’s performative repertoire aimed

at denigrating Clegg.

The findings illustrated in Tables 1 and 2 partially confirm Hypothesis 1b, that Clegg

performs stability and Farage performs crisis. In the first debate, Clegg casts the UK’s

membership of the EU as bringing stability and jobs, as well as granting Britain clout in trade

negotiations by being part of the world’s largest single market economy. However, it is

noteworthy that Clegg’s stability narrative is buttressed by a ‘conditional crisis’ based on the

scenario of a UKIP victory, leading to a British exit from the EU. Clegg communicates this

conditional crisis through claims that the National Health Service would collapse, jobs tied to

foreign investment would be lost, Britain’s clout would shrink in negotiating trade deals, and

the country would become less safe without the European Arrest Warrant. In the second

debate, this conditional crisis comes to the fore of his argument. Clegg attacks Farage

personally, specifically in regards to Farage’s vision to leave the EU as a dangerous fantasy

and a con. Across the two debates, Farage remains on point with his crisis narrative, focusing

on immigration and the loss of national sovereignty. As Table 2 shows, in both debates he

consistently bemoans the free movement of people as open door immigration, the cost of EU

membership, and a loss in the ability to control Britain’s borders due to governance from

Brussels. The results of the word frequency analysis also point out that the policy positions

defended by the two speakers, pro- and anti-EU respectively, remain consistent across the

debates.

23

Logos Pathos Ethos Denigrate Other-10%

0%

10%

20%

30%

40%

50%

60%

70%

17%

49%

6%

17%11%13%

61%

1%

19%

6%

Farage Coding Results

Nigel Farage (LBC) Nigel Farage (BBC)

Figure 3: Qualitative Coding Results

Figure 3 presents the results of the qualitative coding meant to test H1c, which suggested that

Clegg (the non-populist) would rhetorically appeal to ethos and logos and Farage (the

populist) would exhibit pathos and denigration. H1c is, however, only partially confirmed.

While both politicians appeal primarily to pathos (Farage significantly more so than Clegg),

24

Logos Pathos Ethos Denigrate Other-10%

0%

10%

20%

30%

40%

50%

60%

70%

22%

45%

17%11%

5%

18%

37%

14%

28%

3%

Clegg Coding Results

Nick Clegg (LBC) Nick Clegg (BBC)

Clegg uses more references to personal accomplishments (ethos) and facts (logos) than the

Farage, who denigrates his opponent as a consistent strategy.

Clegg’s ethos appeals alluded to the successes of his current government, whereas Farage’s

ethos appeals were largely in connection to his experience in private business as a

commodities broker and his leadership position in the Eurosceptic EP political group Europe

for Freedom and Direct Democracy. Clegg’s logos appeals, especially in the first debate,

related to job creation and were backed by direct quotes from the prominent private sector

managers of Siemens, Hitachi, Nissan, and Ford. For Farage, logos appeals were based on

trade figures, energy prices, and his own party’s estimate of the cost of EU membership.

Clegg’s pathos appeals were most tightly knit to claims that leaving the EU would bring

about a loss of jobs and clout in international negotiations, whereas Farage’s pathos appeals

centered around immigration negatively impacting British workers and diminished national

sovereignty due to EU membership.

Most interestingly, Figure 3 shows that from the first to second debate, Clegg’s denigration

category more than doubles. In the first debate, Clegg primarily attacked ‘UKIP and other

people’ who want to leave the EU. In the second debate, Clegg’s denigration took a much

more personalized scope, attacking Farage personally. These attacks were most often in

reference to Farage’s statement between the two debates that he admired Vladimir Putin,

which was also a prominent word in Table 2. For Farage, denigration was more consistent

and diffuse across the two debates: he often disparaged Clegg together with the other two

national party leaders, framing them as conspiring with multinational corporations and EU

elites. While only a partial confirmation of H1c, the results of the qualitative coding seem to

confirm the adoption hypothesis (H2), that Clegg would embrace a more populist style in the

second debate.

25

Figure 4: MDS Model Figure 5: Hierarchical Clustering

Figures 4 and 5, respectively, visually represent and quantify the change in the politicians’

discourse across the two debates. Together, they help validate the second hypothesis (H2). On

the left, the MDS model represents the politicians’ discourse based on calculations made at

the level of differences in word frequency across the transcripts. The MDS model clearly

illustrates a high similarity between Farage’s discourses in both debates, whereas Clegg’s

discourses are dissimilar both from Farage’s as well as his own across the debates. Figure 5

confirms the findings of the MDS model by using hierarchical clustering, which attempts to

identify clusters between text variables based on their similarity. The results show that

Farage’s transcripts could be clustered together and identified as statistically correlated

(clusters with values of 95 and higher are statistically significant), whereas Clegg’s

transcripts could not be grouped together with statistical certainty.

In sum, Hypothesis 1 tested the theoretical model of populist/technocratic styles via three

measures: degree of formality, performance of crisis, and rhetorical appeals. Of these three

measures, the most accurate demarcation line was drawn by the use of rhetorical appeals.

26

Language formality turned out not to play a consistent role in its association with the

expected populist or technocratic distinction. However, viewed in tandem with the results of

the qualitative analysis, the rather steep decrease in Farage’s formality score from the LBC to

BBC debate may be related to the increase in denigration by Clegg. This suggests that

denigration strategies during a policy debate reduce the quality of substantive information

communicated by the politician being attacked. The performance of crisis, meanwhile, was a

rather ambiguous criterion, as both politicians tried to communicate a sense of impending

alarm, although justified by different reasons. Hypothesis 2 tested the idea of adoption of a

populist style by a mainstream politician who otherwise remained faithful to his policy

positions. Overall, H2 was confirmed. The quantitative methods demonstrate Farage’s

discourse and rhetorical strategies were stable across the debates, albeit a slight deviation in

his formality scores. Clegg, on the contrary, exhibits a clearly observable difference, with the

main change between the debates being the increase in personalized attacks against Farage.

Discussion and ConclusionThis study sought to answer the question: Do mainstream politicians adopt a populist

political style while maintaining a non-populist agenda? The populist style has been defined

here as political communication privileging conflict and crises at the expense of consensus

and substantive facts. The results show that across the debates Nick Clegg, the non-populist

case, reduces his references to substantive information, adopts a personalized attack frame

against his opponent, and conveys a sense of crisis around the potential scenario of Britain

withdrawing from the EU. Although Clegg’s policy position remained staunchly in favor of

EU membership, his political communication style from the first to the second debate shifted

towards one more typical of a populist politician. As populists continue to rise in the polls

and increasingly play the role of kingmaker, mainstream politicians are pressured to adopt

27

elements of the populist style by attacking their opponents’ ethos and attempting to

undermine their legitimacy.

Bale et al. (2010, 413) argue that political parties can incur significant risks by adopting a

populist policy agenda, and the same seems to hold true of mainstream politicians adopting a

populist style. While Clegg was deemed to have lost the first debate with 36% support against

Farage’s 57% (YouGov 2014a), Clegg’s change in style was interpreted unfavorably by

viewers, who gave him only 27% support to Farage’s 68% in the second debate (YouGov

2014b). A host of factors external to the debate help explain Clegg’s defeat: his position in

government, his low popularity from entering a coalition with the Conservatives, his

advocating a status-quo policy position, Farage’s opposition status, and/or a Eurosceptic

British public. However, these external factors do not fully account for why Clegg was

viewed less favorably in the second debate, despite virtually no variation in Farage’s

arguments.

In light of the analysis, the reason posited here is internal to Clegg’s debate performances:

Clegg’s stark contrast in political communication between the two debates contributed to his

coming across as contrived and inauthentic, which ultimately undermined the persuasiveness

of his pro-EU message. Such a change in style may negatively affect trust and, when

combined with a general public dissatisfaction with politics or a particular sense of grievance

regarding a policy area, may undermine the perceived authenticity of mainstream politicians

who adopt the populist style (Liebes 2001; Gilpin et al. 2010).

While the same external factors outlined above similarly contributed to Farage’s victories,

they do not explain why he was perceived to win the second debate by a larger margin than

the first. Given the debates were only two weeks apart, had exactly the same format, and

concerned the same issue, structural factors alone cannot account for the wider spread of

28

Farage’s victory in the second contest. Farage’s discourse and rhetorical appeals across the

debates were surprisingly stable according to both the quantitative and qualitative tests.

Interestingly, Farage was able to win more public support for his position without altering his

message or its delivery. This suggests that one explanatory factor of populists’ success may

be the consistency of their message, and its rhetorical delivery, over time.

Previous work by rhetorical scholars helps explain how such consistency, demonstrated here

at the level of statistical significance, can enhance the persuasiveness of a political argument.

Charteris-Black (2011, 10) stresses that ‘repetition communicates a sense of conviction’ that

can positively contribute to a politician’s ethos. Martin (2014, 98), in a similar vein,

highlights how effective political rhetoric can create ‘feedback loops,’ where earlier

arguments crystallize into enthymematic premises for later ones. Seemingly alluding to

populism, Martin (Ibid.) illustrates the concept of feedback loops by writing, ‘What was once

rhetoric later comes to be ‘common-sense’ premises to routine decisions; what began as an

audacious intervention becomes a coherent discursive frame.’ Supporting this notion, the

findings of the rhetorical analysis demonstrate that Farage did not prioritize the proofs of

factual argumentation (logos) or appeals to his own credibility (ethos) in arguing his anti-EU

position. Rather, the results corroborate earlier findings that Farage’s political communication

is primarily pathos-driven (Crines and Heppell 2016), highlighting the significance of

rhetoric in contemporary politics. If democratic legitimacy can be successfully garnered

through emotional argumentation, there may be less incentive for politicians to ground policy

preferences in verifiable facts. The spreading of a populist style seemingly opens up the door

to post-factual democracy, intimately affecting the electorate’s ability to arrive at informed

political preferences.

This study has argued that Nick Clegg, then Deputy Prime Minister, adopts a populist style

from the first to the second Europe debate in an attempt to ‘fight fire with fire’ against a

29

populist politician. However, in studying only two politicians at two points in time the

conclusions remain tentative. More research is needed, particularly in fragmented multi-party

systems, to further explore the conditions argued here as conducive for mainstream adoption

of the populist style, namely: declining polling numbers for the mainstream politician,

increasing popularity for populist actors, and the saliency of polarizing issues in the media.

Future studies should aim to disentangle these factors and test whether they, independently or

together, can explain mainstream adoption of the populist style. Despite its limitations, this

study provides further evidence that although British politicians have been argued to resist

the conflict-oriented political style witnessed by established politicians in other national

contexts (Brants and Voltmer 2011, 13), the populist style has penetrated the mainstream in

British politics.

References Akkerman, T. (2012) ‘Comparing radical right parties in government: Immigration and integration policies in nine countries (1996–2010)’, West European Politics, 35:3, 511-529.

Alexander, J. (2011) Performance and Power (Cambridge: Polity).

Aslanidis, P. (2015) ‘Is populism an ideology? A refutation and a new perspective,’ Political Studies (published online ahead of print 5 October). Available at: http://onlinelibrary.wiley.com/doi/10.1111/1467-9248.12224/abstract (Accessed 5 October 2015).

Atkins, J (2011) Justifying New Labour Policy. (Basingstoke: Palgrave Macmillan).

Atkins, J. and Finlayson, A., (2013) ‘…A 40-year-old black man made the point to me’: Everyday knowledge and the performance of leadership in contemporary British politics’, Political Studies, 61:1, 161-177.

Atkins, J. and Turnbull, N. (2014) ‘Conclusion: Rhetoric, British identity and interdisciplinarity’, in J. Atkins, A. Finlayson, J. Martin and N. Turnbull (eds) Rhetoric in British Politics and Society (Basingstoke: Palgrave Macmillan), 173-184.

Bale, T., Green-Pedersen, C., Krouwel, A., Luther K.R. and Sitter, N. (2010) ‘If you can't beat them, join them? Explaining Social Democratic responses to the challenge from the populist radical right in western Europe’, Political Studies 58:3, 410-426.

Brants, K. and Voltmer, K. (2011) ‘Introduction: mediatization and de-centralization in political communication’, in K. Brants and K. Voltmer (eds) Political Communication in Postmodern Democracy: Challenging the Primacy of Politics (Basingstoke: Palgrave Macmillan), 1-18.

Budd, R., Thorp, R.K. and Donohew, L. (1967) Content Analysis of Communications (New York: Macmillan).

30

Canovan, M. (1984) ‘‘People’, politicians and populism’, Government and Opposition, 19:03, 312-327.

Charteris-Black, J. (2011). Politicians and Rhetoric: The Persuasive Power of Metaphor (Basingstoke: Palgrave Macmillan).

Creswell, J. (2014) Research Design (Thousand Oaks, CA: Sage).

Crines, A. S. (2013a) ‘The rhetoric of the coalition: Governing in the national interest?’, Representation, 49:2, 207-218.

Crines, A. S. (2013b) ‘An analysis of George Galloway’s oratorical and rhetorical impact,’ Politics, 33:2, 81-90.

Crines, A. S. and Hayton, R. (2015)  Labour orators from Bevan to Miliband. Manchester University Press.

Crines, A. S. and Heppell, T. (2016) ‘Rhetorical style and issue emphasis within the conference speeches of UKIP’s Nigel Farage 2010–2014’, British Politics’, (published online ahead of print 5 October). Available at: http://link.springer.com/article/10.1057/bp.2016.2 (Accessed 5 October 2015).

Dommett, K. (2014) ‘Rhetoric and party politics—Looking beyond the leader’, In in J. Atkins, A. Finlayson, J. Martin and N. Turnbull (eds) Rhetoric in British Politics and Society (Basingstoke: Palgrave Macmillan), 73-86.

Dutceac Segesten, A. and Bossetta, M. (2016) ‘A typology of political participation online: How citizens used Twitter to mobilize during the 2015 British general elections’, Information, Communication & Society, (published online ahead of print 6 November). Avaiable at: http://www.tandfonline.com/doi/full/10.1080/1369118X.2016.1252413

Gilpin, D., Palazzolo, E. and Brody, N. (2010) ‘Socially mediated authenticity,’ Journal of Communication Management, 14:3, 258–278.

De Lange, S.L. (2012) ‘New alliances: Why mainstream parties govern with radical right-wing populist parties’, Political Studies, 60:4, 899-918.

Finlayson, A. and Martin, J. (2008) ‘‘It ain’t what you say…’: British political studies and the analysis of speech and rhetoric’, British Politics, 3:4, 445-464.

Finlayson, A. (2012) ‘Rhetoric and the political theory of ideologies’, Political Studies, 60:4, 751-767.

Fry, E. (1997) 1000 Instant Words (Lincolnwood, IL: Contemporary Books).

Hay, C. (1996) ‘Narrating crisis: The discursive construction of the winter of discontent’, Sociology, 30:2, 253-277.

Hayton, R., & Crines, A. S. (2015) Conservative Orators from Baldwin to Cameron. Manchester University Press.

Heylighen, F. and Dewaele, J.M. (2002) ‘Variation in the contextuality of language: An empirical measure’, Foundations of Science, 7:3, 293-340.

Hooghe, L. and Marks, G. (2005) ‘Calculation, community and cues public opinion on European integration’, European Union Politics, 6:4, 419-443.

Jagers, J. and Walgrave, S. (2007) ‘Populism as political communication style: An empirical study of political parties discourse in Belgium’, European Journal of Political Research, 46:3, 319-45.

Kaufman, L. and Rousseeuw, P.J. (2005) Finding Groups in Data: An Introduction to Cluster Analysis (Hoboken: John Wiley & Sons).

Kazin, M. (1998) The Populist Persuasion: An American History (Ithaca: Cornell University Press).

31

Krebs, R.R. and Jackson, P.T. (2007) ‘Twisting tongues and twisting arms: The power of political rhetoric’, European Journal of International Relations, 13:1, 35-66.

Kriesi, H. (2014) ‘The populist challenge’, West European Politics, 37:2, 361-378.

Kriesi, H. and Pappas, T.S. (2015) ‘Populism in Europe during crisis: An introduction’, in H. Kriesi and T.S. Pappas (eds)  European Populism in the Shadow of the Great Recession. (Colchester: ECPR Press), 1-22.

Laclau, E. (2005) On Populist Reason (London: Verso).

Liebes, T. (2001) ‘Look me straight in the eye: The political discourse of authenticity, spontaneity, and sincerity’, The Communication Review, 4:4, 499-510.

Martin, J. (2014) Politics and Rhetoric: A Critical Introduction (New York: Routledge).

Martin, J. (2015) ‘Situating speech: A rhetorical approach to political strategy’, Political Studies, 63:1, 25-42.

Marcus, G.R. (2000) ‘Emotions in politics’, Annual Review of Political Science, 3:1, 221-250.

Mazzoleni, G. (2014) ‘Mediatization and political populism’, in: F. Esser and J. Strömbäck (eds), Mediatization of Politics: Understanding the Transformation of Western Democracies (Basingstoke: Palgrave Macmillan), 3-30.

Moffitt, B. (2015) ‘How to perform crisis: A model for understanding the key role of crisis in contemporary populism’, Government and Opposition, 50:2, 189-217.

Moffitt, B. (2016) The Global Rise of Populism (Stanford: Stanford University Press).

Moon, D.S. (2014) ‘Rhetoric and devolution – Time and space in Welsh Labour rhetoric’, in J. Atkins, A. Finlayson, J. Martin and N. Turnbull (eds) Rhetoric in British Politics and Society (Basingstoke: Palgrave Macmillan), 87-100.

Mudde, C. (2004) ‘The populist zeitgeist’, Government and Opposition, 39:4, 542-563.

Mudde, C. (2013) ‘Three decades of populist radical right parties in Western Europe: So what?’, European Journal of Political Research, 52:1, 1-19.

Pappas, T. (2016) ‘Modern populism: Research advances, conceptual and methodological pitfalls, and the minimal definition’, in: Oxford Research Encyclopedia of Politics. Available at: http://politics.oxfordre.com/view/10.1093/acrefore/9780190228637.001.0001/acrefore-9780190228637-e-17 (accessed 14 October 2016).

Parry, K. and Richardson, K. (2011) ‘Political imagery in the British general election of 2010: The curious case of Nick Clegg’, The British Journal of Politics & International Relations, 13:4, 474-89.

Reisigl, M. (2008) ‘Analyzing political rhetoric’, in: R. Wodak and M. Krzyżanowski (eds), Qualitative Discourse Analysis in the Social Sciences (Basingstoke: Palgrave Macmillan), 96-120.

Rinker, T.W. (2013) qdap: Quantitative discourse analysis R package 2.2.4. (University at Buffalo. Buffalo, New York). Available at: http://github.com/trinker/qdap

Rolfe, M. (2016) The Reinvention of Populist Rhetoric in The Digital Age: Insiders & Outsiders in Democratic Politics. (Singapore: Springer Nature).

Rooduijn, M. and Pauwels, T. (2011) ‘Measuring populism: Comparing two methods of content analysis’, West European Politics, 34:6, 1272-1283.

Rooduijn, M. (2014) ‘The nucleus of populism: In search of the lowest common denominator’, Government and Opposition, 49:4, 573-99.

32

Rooduijn, M., De Lange, S.L. and Van Der Brug, W. (2014) ’A populist Zeitgeist? Programmatic contagion by populist arties in Western Europe’, Party Politics, 20:4, 563-575.

Schumacher, G. and Van Kersbergen, K. (2016) ‘Do mainstream parties adapt to the welfare chauvinism of populist parties?’, Party Politics, 22:3, 300-312.

Stanley, B. (2008) ‘The thin ideology of populism’, Journal of Political Ideologies, 13:1, 95-110.

Suzuki, R. and Shimodaira, H. (2015) pvclust: Hierarchical clustering with p-values via multiscape bootstrap resampling R package version 2.0-0. Available at: https://CRAN.R-project.org/package=pvclust

Tournier‐Sol, K. (2015) ‘Reworking the Eurosceptic and conservative traditions into a populist narrative: UKIP's winning formula’, Journal of Common Market Studies, 53:1,140-56.

Van Spanje, J. (2010) ‘Contagious parties, anti-immigration parties and their impact on other parties’ immigration stances in contemporary Western Europe’, Party Politics, 16:5, 563-586.

Vasilopoulou, S. Halikioloulou, D. and Exadaktylos, T. (2014) ‘Greece in crisis: Austerity, populism, and the politics of blame’, Journal of Common Market Studies 52:2, 388-402.

YouGov (2014a) YouGov / Sun Survey Results [online] YouGov. Available at: http://d25d2506sfb94s.cloudfront.net/cumulus_uploads/document/jf8oxz4fur/YouGov_The%20Sun%20140326%20Clegg_Farage_EU_Debate.pdf (accessed 14 October 2016).

YouGov (2014b) YouGov / Sun Survey Results [online] YouGov. Available at: http://d25d2506sfb94s.cloudfront.net/cumulus_uploads/document/ra091t82t7/Full_Results_Debate_140402.pdf (accessed 14 October 2016).

Washbourne, N. (2013) ‘More than Cleggmania? The Celebrity Politician, Presidentialization and the UK 2010 Televised Leader Debates’, in M. Ekström and A. Tolson (eds) Media Talk and Political Elections in Europe and America (Basingstong: Palgrave Macmilan), 113-132.

33